Henry Rzepa's Blog Chemistry with a twist

December 5, 2009

The nature of the C≡S Triple bond: Part 2

In my first post on this theme, an ELF (Electron localization function) analysis of the bonding in the molecule HO-S≡C-H (DOI: 10.1002/anie.200903969) was presented. This analysis identified a lone pair of electrons localized on the carbon (integrating in fact to almost exactly 2.0) in addition to electrons in the CC region. This picture seems to indicate that the triple bond splits into two well defined regions of electron density (synaptic basins). In a comment to this post, I also pointed out that an NBO analysis showed a large interaction energy between the carbon lone pair and the S-O σ* orbital, characteristic of anomeric effects (in eg sugars). This latter observation gives us a handle on possibly tweaking the effect. Thus if the S-O σ* orbital can be made a better electron acceptor, then its interaction with the lone pair could be enhanced.

Accordingly, the analysis has been repeated for H-C≡S-OTf (OTf = triflate = trifluoromethane sulfonate), since the triflate would be expected to increase significantly the electron accepting properties of the S-O bond.

ELF analysis for H-C≡S-OTf

ELF analysis for H-C≡S-OTf. Click for 3D model

One dramatic change has indeed occurred. Previously, a well-defined ELF disynaptic basin had been identified in the S-O region, with an integration of 1.12e. If the OH group is replaced by OTf, this disynaptic basin can no longer be located. The electrons have instead moved into sulfur lone pairs, and the S-CF3 bond, which is an expected consequence of the greater electronegativity of the triflate group. Point 15 (the S=C region) integrates to 2.56e (compared with 2.36 for the OH analogue), and the carbon lone pair decreases from 2.01 to 1.86e.

Taken as a whole, these changes suggest that the CS bond has gotten stronger, resulting from transfer of electron density from the non-bonding carbon lone pair, to the CS bond itself. Indeed, its length is now 1.492Å, a significant shrinkage compared to 1.544Å for the OH parent (B3LYP/cc-pVTZ). Likewise, the C-S vibrational stretch of 1381 cm-1 for the OTf derivative is an increase over 1215 cm-1 for the OH system and 1304 cm-1 for diatomic CS itself (B3LYP/cc-pVTZ).

These results suggest that the ELF procedure, combined with the insight from the NBO analysis, can be used as a tool to rationally design a variation to the original molecule which does appear to enhance the triple bond character of the CS region, and to fulfil further the ambition of the original article by Schreiner and co-workers. As a triflate, it may even be susceptible to a simple preparation from the alcohol parent! Anyone up for it?

It is also worth noting that the above system is headed off towards HC≡S+, the thioacylium cation (although crystal structures of the acylium ion are known, none have been reported for the thioacylium ion). Both N≡N and C≡O contract their bond lengths when protonated, so it should be no great surprise to find that CS does so as well (1.476Å, ν 1543 cm-1).


  1. As a follow-up to the post, the HC≡S-OTf system was noted to have a (gas phase) dipole moment of 5.6D, the result of shifting charge from the carbon lone pair to the S-OTf region. This dipole is the value at which solvation starts to have a discernible influence. Thus re-optimizing the geometry (B3LYP/cc-pVTZ/SCRF, DOI: 10042/to-3155) for a relatively non-polar solvent (benzene) is sufficient to induce the following changes: the dipole moment increases to 9.3D, the CS bond length decreases to 1.482Å, and the νCS stretch increases to 1424 cm-1. For a more polar solvent (dichloromethane, DOI: 10042/to-3157), the values continue this trend towards HCS+: 13.3D, 1.476Å and 1446 cm-1.

    Comment by Henry Rzepa — December 5, 2009 @ 6:40 pm

  2. […] Henry Rzepa Chemistry with a twist « The nature of the C≡S Triple bond: Part 2 […]

    Pingback by The nature of the C≡S triple bond: part 3. « Henry Rzepa — December 6, 2009 @ 11:57 am

  3. […] degree of SC triple character (HCS+) has around 3.8 in the same region, and perhaps a better model (TfOSCH, for which the carbon, like here, has a carbene lone pair) has 2.6. Each carbon also has the […]

    Pingback by Contriving aromaticity from S≡C Triple bonds « Henry Rzepa — January 1, 2010 @ 1:03 pm

RSS feed for comments on this post. TrackBack URL

Leave a comment

Powered by WordPress